This test passively measures the amount of data used on the user's broadband connection. This test does not inspect or record the contents of packets; it only records volumes. The reporting interval is configurable, and defaults to one hour. It can go as low as once per minute. All traffic volumes are captured at the frame level, meaning that they include Ethernet, IP and TCP/UDP header overheads.
The test captures traffic volumes from the WAN interface only. In other words, it does not include LAN-to-LAN traffic.
The precise mechanism used to record data usage volumes varies by hardware platform. In some devices a managed hardware Ethernet switch is present; in these cases the WAN port counters are simply interrogated and recorded (the CPU does not even see the traffic). In other devices only a software switch is present, and in these cases the counters from the software bridge in the Linux kernel are taken.
The key metrics captured by this test are:
- Downstream bytes received over the WAN in this reporting interval
- Upstream bytes sent over the WAN in this reporting interval
- The start time and end time of the reporting interval
These metrics are commonly used by our ISP partners to validate their own counters and/or generate thorough traffic consumption reports. They can also use this capability as a supplementary data source to improve the accuracy of their billing.
We run closed audits to help ISPs spot any anomalies in their billing systems that record customer’s usage. How does this work? Well, we deploy the Data usage test onto the routers of participating households. The test tracks the internet usage of those households. We can then compare this to the data usage calculated by the ISPs billing system. If we see any anomalies, we can point them out so the ISP can investigate and rectify them.